test_ChatAPI.py

import json
from pickle import TRUE
from train import all_words, dataset, intents
import torch
from model import NeuralNet
from nltk_utils import bag_of_words, tokenize
import random

class ChatAPIs():


    def chat (self, sentence):
        #if GPU available
        device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

        #Load sentiment data
        with open('intents.json', 'r') as f:
            intents = json.load(f)

        #Load saved model
        FILE = "trained.pth"
        data = torch.load(FILE)

        input_size = data["input_size"]
        hidden_size = data["hidden_size"]
        output_size = data["output_size"]
        all_words = data["all_words"]
        tags = data["tags"]
        model_state = data["model_state"]

        model = NeuralNet(input_size, hidden_size, output_size).to(device)
        model.load_state_dict(model_state)
        model.eval()

        bot_name = "Ghost"
        
        sentence = tokenize(sentence)
        X = bag_of_words(sentence, all_words)
        X = X.reshape(1, X.shape[0])
        X = torch.from_numpy(X).to(device)

        output = model(X)
        _, predicted = torch.max(output, dim=1)
        tag = tags[predicted.item()]

        probs = torch.softmax(output, dim=1)
        prob = probs[0][predicted.item()]

        #print (prob.item())
        if prob.item() > 0.40: #We found a match with over 70% probability
            for intent in intents["intents"]:
                if tag == intent["tag"]:
                    out = (f"{random.choice(intent['responses'])}")
        else:
            out = ("Nothing matched...")
        return out


    #def chat( sentence):
    #    return sentence

def tchat(text):
    api = ChatAPIs()
    print(api.chat(text))

#tchat("hello")  # This works now

def test_module():
    assert (TRUE)

def test_chat():
    assert tchat("Hello") != ""